# Workshop Feedback Form — Reading the Results

> Three numbers, two files, one meeting. Reading takes 10 minutes per event.
> Use with `workshop-feedback-form-guide.md`.

---

## The 60-second read

1. **Recommendation score at a glance.** The median of the 0-10 scores. 8.5 or higher: the workshop sells itself to returners. Below 6: read the item means before the testimonials — something structural is wrong.
2. **Quotable testimonial count.** Sentences that pass the specific test — a named action, a before/after, a number. Two strong quotes per event is a good harvest; five is a great one.
3. **Top demanded topic.** The answer that repeats. One repeat is a candidate; three repeats is your next event, already validated by the asking.

---

## Item-by-item interpretation

| Item | Construct | Low (≤ 2.5) means | First action |
|---|---|---|---|
| 1 | Promise kept | The page or pitch oversold the content | Rewrite the description, not the workshop |
| 2 | Pace & length | Material crammed or stretched | Re-time the agenda blocks |
| 3 | Working, not listening | The workshop drifted into lecture | Cut one content block, add one exercise |
| 4 | Standalone usefulness | The useful part felt like a teaser for something paid | Move the teaser's payoff into the room |

Item 4 is the trust item for independent facilitators: the moment the room senses the workshop is a funnel, scores and testimonials decay together. Useful-without-upsell is a feature — protect it.

---

## Reading the recommendation score

Compute NPS if you want the industry number: promoters (9-10) minus detractors (0-6), as percentages. But at workshop scale — under 30 responses — the percentage amplifies noise: one detractor moves NPS by three or more points.

At small scale, read the raw distribution instead:

- A pile of 9-10s **and** a pile of 0-3s means the workshop landed for one audience and missed another — check whether the two piles describe different job roles.
- A tight 6-8 band means solid but not remarkable: safe to keep running, not yet safe to raise the price.

---

## Coding the two open questions

**Testimonials → the quote file.** Keep every permissioned quote in one running document, tagged with event name and date. When writing a workshop page, use quotes that name an action ("I rewrote my kickoff agenda"), not quotes that name you ("the facilitator was great"). The first kind convinces strangers; the second convinces no one.

**Demand → the programming calendar.** Cluster the answers into topics, then count votes. Publish the top three back to the attendees of that event ("you asked — here is the next quarter") — closing the loop is what converts one-time attendees into a following.

---

## 3 known failure modes of reading the results

1. **The best quote is anonymous.** You cannot use it. Rescue: the permission checkbox is on the testimonial card for exactly this reason — point at it while handing cards out.
2. **One detractor's item scores hijack the redesign.** A single hostile card is one person's experience. Rescue: cluster the demands and read the testimonials before touching the agenda.
3. **The demand list dies in a drawer.** Collected and never acted on, attendees learn that the form is theater. Rescue: publish the next-quarter topic list back to the same email list — the loop has to be visible to keep working.